The Optical

Scintillation by

Extraterrestrial

Refractors

Project

Tucson 03/19/2010 Marc MONIEZ, IN2P3, CNRS

Search for Galactic hidden gas

A&A 412, 105-120 (2003)(astro-ph/0302460)

• Cold (10K) => no emission. Transparent medium.

• In the thick disk or/and in the halo

• Average column density toward LMC

OSER project: measure the last unknown contribution to the baryonic hidden matter: cold H2 (+He)

250g/m2 <=> columnof 3m H2 (normal cond.)

• Fractal structure: covers ~1% of the sky.Clumpuscules ~10 AU(Pfenniger & Combes 1994) ~300m H2 over 1% of the sky

These clouds refract light

• Extra optical path due to H2 medium Varies from 0 (99% sky) to ~80 000 (1%) @

=500nm

• If the medium has column density fluctuations (turbulences) of order of a few 10-6 then wavefront distorsions may be detectable

Scintillation through a diffusive screenPropagation of distorted wave surface driven by:

Fresnel diffraction + « global » refraction

Pattern moves at the

speed of the screen

Mo

no

ch

rom

ati

c

=2

.18

Po

lyc

hro

ma

tic

(Ks

pas

sb

an

d)

Point source Extended source

m=1.08

m=0.76

m=0.23

m=0.23m = I/I = modulation index

Illumination in Ks by a K0V star@8kpc (mV=20.4) through a cloud@160pc with Rdiff =150km

Simulation towards B68

Simulated light curve

Distance scales5 distance scales are found in the speckle pattern• Diffusion radius Rdiff

• separation such that: [(r+Rdiff)-(r)] = 1 radian• Characterizes the turbulence

• Fresnel radius RF

• scale of Fresnel diffraction ~103 km @=1 for gas@1kpc

• Refraction radius Rref

• diffractive spot of Rdiff patches ~ 2RF2/Rdiff

• Larger scale structures of the diffusive gaz can play a role if focusing/defocusing configurations happen

• Projected source size RSspeckle from a pointlike source isconvoluted by the source projected profile

Rdiff : Statistical characterization of a stochastic screen

Size of domain where(phase)= 1 radian• i.e. (at = 500 nm)(column density nl) = 1.8x1018 molecules/cm2

• This corresponds to- nl/nl ~ 10-6 for disk/halo clumpuscule- nl/nl ~ 10-4 for Bok globule (NTT search)

Modulation indexEssentially depends onRS/Rref

-> not on the details of the power spectrum of the fluctuations

RS/Rref

I/I

z1 is the cloud-source distance

Scintillation of Sun@10kpc

through a cloud with Rdiff=1000km

at = 1 m

Time scale

If Rdiff < Rref, then Rref is the largest scale and :

Wherez0 is the distance to the cloudVT is the relative speed of the cloud w/r to the l.o.s.

Signature of scintillation• Stochastic light-curve (not random)

– Autocorrelation (power spectrum)– Characteristic times (10’s minutes)– Modulation index can be as high as 5%

• decreases with star radius• depends on cloud structure

• Signatures of a propagation effect– Chromaticity (optical wavelengths)

• Long time-scale variations (10’s min.) ~ achromatic• Short time-scale variations (~ min.) strongly change with

– Correlation between light-curves obtained with 2 telescopes decreases with their distance

Fore and back-grounds

• Atmospheric turbulencePrism effects, image dispersion, BUT I/I < 1% at any time scale in a

big telescope

BECAUSE speckle with 3cm length scale is averaged in a >1m aperture

• High altitude cirrusesWould induce easy-to-detect collective absorption on neighbour

stars. Scintillation by a 10AU structure affect one only star.

• Gas at ~10pcScintillation would also occur on the biggest stars

• Intrinsic variabilityRare at this time scale and only with special stars (UV Ceti, flaring Wolf-Rayet)

Optical depth due to halo gas towards LMC

scintillation = 10-2 x x S

Where• 10-2 is the max. surface coverage of the fractal structures • is the fraction of halo into molecular gas• S depends on the structuration… Unknown

Telescope > 2 metersFast readout Camera

2 camerasWide field

Requirements for detection towards LMC

• Assuming Rdiff = 1000km (10 AU clumpuscules)• 5% modulation@500nm => rs < rA5 (105/deg2)

Smaller than A5 type in LMC => MV~20.5 Characteristic time ~ 1 min. => few sec. exposures Photometric precision required ~1%

Dead-time < few sec. => B and R partially correlated => Optical depth probably small =>

What will we learn from detection or non-detection?

• Expect 1000x fluctuating light-curves (>5%) when monitoring 105 stars if column density fluctuations > 10-6

within 1000km• If detection

– Get details on the clumpuscule (structure, column density -> mass) through modelling (reverse problem)

– Measure contribution to galactic hidden matter

• If no detection– Get max. contribution of clumpuscules as a function of their

structuration parameter Rdiff (fluctuations of column density)

• 9599 stars monitored• ~ 1000 exposures/night• Search for ~10%

variability on the 933 best measured stars

• Signal if nl)/(nl) ~ 10-4 per ~1000 km (nl = column density)

• Mainly test for back-ground and feasibility

Test towards Bok globule B68NTT IR (2 nights in june 2006)

Test towards Bok globule B68

NTT IR (2 nights)

one fluctuating star?(other than known artifacts)

⁄⁄

Time (103s)

Conclusion: optical interstellar scintillation with LSST

• LSST has the capacity for this search– Large enough (diameter and field)– Fast readout (2s) -> allows fast sampling

• But current sampling strategy does not fit– Need few hours series of short exposures– But no need for regular long time sampling (contrary to

microlensing or SN searches)

• Complementary synchronized observations for– test of chromaticity– decorrelation with distant simultaneous observations

complements

For the future…

A network of distant telescopes• Would allow to decorrelate scintillation from

interstellar clouds and atmospheric effects• Snapshot of interferometric pattern + follow-up

Simultaneous Rdiff and VT measurements

=> positions and dynamics of the cloudsPlus structuration of the clouds (inverse problem)

Expected difficulties, cures• Blending (crowded field)=> differential photometry• Delicate analysis

– Detect and Subtract collective effects– Search for a not well defined signal

• VIRGO robust filtering techniques (short duration signal)• Autocorrelation function (long duration signal)• Time power spectrum, essential tool for the inversion

problem(as in radio-astronomy)

• If interesting event => complementary observations (large telescope photometry, spectroscopy, synchronized telescopes…)